65results about How to "Improve defect detection performance" patented technology

The invention relates to a mobile phone casing defect detecting method based on depth learning. The method comprises the steps that (1) the image of a mobile phone casing to be detected is acquired and pre-processed; and (2) the pre-processed image is input into a pre-trained defect detection model for defect detecting to acquire the position of a defect on the mobile phone casing, and the confidence of the position as the defect is provided. The defect detection model is a depth network based on depth learning, and comprises a feature extraction network and a classifier and regression device network, wherein the feature extraction network and the classifier and regression device network are in successive cascade. The feature extraction network carries out feature extraction on the pre-processed image to acquire a feature image. The classifier and regression device network classifies and regresses the feature image to acquire the defect position and the confidence of the mobile phone casing. Compared with the prior art, the method provided by the invention has the advantages of high detection precision and accurate and reliable detection result.

An imaging system for inspection of a sample includes an illumination module, which irradiates a surface of the sample with pulsed optical radiation. A mechanical scanner translates at least one of the sample and part of the imaging system so as to scan an area irradiated by the pulsed optical radiation over the surface in order to irradiate successive, partially overlapping frames on the surface by respective successive pulses of the pulsed radiation. A collection module collects the optical radiation scattered from the surface so as to capture a sequence of images of the irradiated frames. A system controller varies a configuration of the imaging system in alternation between at least first and second different optical configurations in synchronization with the pulsed optical radiation.

The invention relates to a method for detecting the welding quality of a hot-melt even joint of a polyethylene pipeline and a detection device thereof. The aim of the invention is that the provided detection method and the detection device have the characteristics of high sensitivity, high detection precision, high reliability, high detection efficiency and low misoperation rate. The technical proposal is a method for the coupling focusing ultrasonic detection of the hot-melt even joint of the polyethylene pipeline, which comprises the following steps: (1) a coupling agent liquid is positioned on the outer surface of a pipeline joint fusion region; (2) an ultrasonic probe adopting a spherical curved wafer is immersed into the coupling agent liquid in order that the focus is positioned in the centerline of the pipeline joint fusion region; and (3) the ultrasonic probe does linear scanning motion along the diameter direction of the pipeline and does uniform scanning motion along the circumferential direction of the pipeline; and obtained signal data is processed by a computer to form a C scanning picture. The ultrasonic detection device comprises a coupling receiving device, the ultrasonic probe adopting the spherical curved wafer, a mechanical scanning device, a signal processing and controlling system, and the computer.

The invention discloses a product defect detection method and device, electronic equipment and a storage medium, and relates to the fields of computer vision, image processing, deep learning and the like. According to the specific implementation scheme, the method includes obtaining a to-be-detected image of a target product; determining a color feature difference value between the to-be-detectedimage and a template image according to the color feature value of the to-be-detected image and a pre-stored color feature value of the template image; taking the color feature difference value as a feature value of the first channel to obtain an input image comprising the first channel; and obtaining defect information of the target product according to the input image and the target detection model. According to the embodiment of the invention, the defect detection effect on products with small sizes and weak texture features in an industrial scene can be improved.

The invention provides an improved method and device for inspecting heat exchanger tubes from within the tube inside diameter that overcomes the shortcomings of the prior art. It adapts a guided-wave probe approach that makes use of a torsional wave mode instead of a longitudinal wave node disclosed in the prior art. The torsional wave mode has many advantages over the longitudinal wave mode for detecting defects. When energized by suitable instrumentation, the probe is caused to generate a torsional mode signal that is transmitted to the heat exchanger tube from the waveguide tube. When reflected signals from defects in the heat exchanger tube walls are returned to the inspection opening end of the heat exchanger tube, the reflected defect signals are transmitted to the probe waveguide tube for amplification, detection and characterization of the reflected signal.

The invention provides an outer package defect detection method. The method comprises the steps of obtaining an outer package picture as a data set; performing data preprocessing on the data set; setting the size of a defect target candidate box through a clustering algorithm; inputting the preprocessed data set into a Cascade R-CNN model, and performing region-of-interest extraction on the data set; normalizing the size of the feature map of the region of interest; in the Cascade R-CNN model, adopting a pre-trained ResNet-50 with a deformable convolution v2 as a backbone network, carrying outconvolution, and extracting features; and S44, framing out defects and marking defect categories. The outer package defect detection method is simple in calculation process and capable of rapidly positioning and recognizing defects, the FPN feature extraction method is adopted, features with different resolutions are combined, the combined features are used for enhancing multi-level features in FPN, and recognition is more accurate.

The invention discloses an ultrasonic automatic detection method capable of considering workpiece clamping error correction. Under the condition that a clamping error exists in the workpiece, an industrial robot carries an ultrasonic probe to perform C scanning detection route correction to realize clamping error compensation. The method particularly comprises the following steps: identifying the actual clamping error of the workpiece through a change rule of water sound distance along with time in the rotating process of the workpiece, calculating a route transformation matrix from a workpiece coordinate system to a world coordinate system, and performing route transformation calculation on a scanning target route under the workpiece coordinate system to obtain a moving route of a probe under the world coordinate system to obtain the corrected probe moving route.

The invention discloses a bearing roller chamfered surface defect detection method based on machine vision. Aiming at the chamfering surface defect problem of the bearing roller in the production andtransportation process, a chamfering surface image of the bearing roller is collected, a chamfering surface missing part detection model is established by utilizing a deep learning algorithm based ona deep convolutional neural network, and rapid detection and position positioning of the chamfering surface defect are realized. According to the method, the deep learning model and the machine visionalgorithm are used for detecting the bearing roller chamfer defects in real time, the method has the advantages of being rapid in detection, accurate in positioning and high in recognition accuracy,a traditional manual detection method can be replaced, and the requirement for roller defect detection automation is met.

The invention relates to a microwave thermal imaging non-destructive detection system based on matrix decomposition, comprising a microwave excitation device and a thermal imaging monitoring device; the microwave excitation device includes a microwave signal generator, a microwave signal amplifier, a microwave excitation sensor and a wave-absorbing material; the thermal imaging monitoring device includes A thermal imager, a data acquisition and processing device; the data acquisition and processing device is respectively connected with the microwave signal generator and the thermal imager, so as to obtain the temperature change data of the measured object. The microwave thermal imaging non-destructive testing system based on matrix decomposition can solve the problem of rapid imaging and separation of different types of defects, highlight the feature extraction of thermal spatial patterns in the defect range, and solve the problem that adjacent defects cannot be located and separated due to temperature aliasing and the damage area of ​​subsurface defects It is difficult to effectively quantify the problem; solve the problem of automatic separation of defects, and can quickly image the impact defects and damage areas of different energies.

The invention discloses a method for correcting and compensating ultrasonic C scanning paths under the circumstance that eccentric clamping between the center of a rotation body workpiece and a turnplate rotating shaft occurs when the ultrasonic longitudinal wave vertical incidence rotation body workpiece is adopted to detect inside defects. The method includes: detecting actual deviation of the workpiece through a fixed ultrasonic probe; calculating a path transformation matrix from a workpiece coordinate system to a world coordinate system; subjecting a scanning target path of the workpiece under the workpiece coordinate system to matrix transformation calculation to acquire moving paths of the probe under the world coordinate system.

The invention discloses an addressable test chip and a test method thereof. The test chip consists of an addressable circuit, a plurality of test frameworks and a plurality of test structure groups which are placed among the test frameworks, wherein the addressable circuit consists of a peripheral address decoding circuit and a plurality of switch circuits. The peripheral address decoding circuit consists of a row address decoding circuit, a column address decoding circuit and a pin selection decoding circuit. All the input ends of the row address decoding circuit, the column address decoding circuit and the pin selection decoding circuit are connected with a PAD set, and the output ends are connected with the switch circuits; the switch circuits are also connected to a PAD signal location by the plurality of signal lines; the switch circuits are connected with the test structure groups by the test frameworks. During test, the test frameworks are selected by row address signals and column address signals, signal selection pins are selected by pin selection signals, and test structures are tested by test signals. By means of the pin selection circuit, the addressable test chip and the test method can test the test structure group with a plurality of pins.

The invention provides a detection method for surface defects. The detection method for surface defects comprises the following steps: the surface frequency spectrum of a metal machining part to be detected and the surface frequency spectrum of a metal machining part free of surface defects are subtracted, so as to obtain a frequency spectrum of defects; the frequency spectrum of defects is actually the frequency spectrum on which the frequency spectrum of the surface background of the metal machining part to be detected is removed; therefore, when the inverse Fourier transform is conducted on the frequency spectrum of the defects, the surface image on which the surface background of the metal machining part to be detected is removed is obtained; the surface image is displayed, the situation that whether the surface of the metal machining part to be detected has defects or not and the distribution of the defects can be clearly distinguished. Therefore, the detection device and the detection method for the surface defects can weaken the interferences of the surface background of the metal machining part to the detection process; the effect of the defect detection is improved and the speed of defect detection is increased. Moreover, the invention further provides a detection device for the surface defects.

The invention discloses an infrared thermal image analysis method for nondestructive testing of defects of a composite material, which comprises the following steps of (1) data acquisition: acquiringthermal image data of internal defects of a polymer composite material by using a thermal infrared imager, including two processes of pulse heating of a sample and capture of a thermal image on the surface of the sample, (2) data preprocessing: preprocessing the original thermal image data acquired by the sensor, wherein the preprocessing comprises the following three parts: a) selecting an effective area of the image, b) expanding the data, and c) centralizing the data, (3) constructing a defect detection thermal image data analysis model OLPPT: calculating an embedded matrix Y of a feature space, (4) carrying out OLPPT model result visualization: carrying out embedded projection y visualization, and (5) quantitatively evaluating the effect of the OLPPT model: calculating a signal-to-noise ratio index.

The invention provides a tire flaw detection method and model based on a self-attention mechanism and dual-field self-adaption, and the method comprises the steps: cutting a tire X-ray image in a model training stage, and inputting the cut tire X-ray image into a tire flaw detection feature extraction module composed of ResNet50 and a VIT model; carrying out marginal probability distribution adaptation on the obtained source domain features and target domain features by using MK-MMD, and carrying out conditional probability distribution adaptation on the same category of samples in the source domain and the target domain by using LMMD; and a defect detection result is obtained through a defect classifier subsequently, and the model is trained in a back propagation mode. In the flaw detection stage, after a tire X-ray image is cut, the tire X-ray image is directly input into the tire flaw detection feature extraction module obtained in the model training stage; and inputting the obtained features into a defect classifier obtained in the model training stage to output a defect detection result. According to the invention, domain offset between images sampled by different X-ray machines can be overcome, and the method has good performance in tire flaw detection with deficient data.

The present invention provides a defect inspection method and device for a display panel. The defect inspection method comprises: A) obtaining an edge image of the display panel and obtaining a grayscale value of each pixel of the edge image; B) selecting a specific area in the edge image and obtaining a grayscale value of each pixel in the specific area; C) obtaining an average grayscale value of all pixels in the specific area; D) executing binarization for the grayscale value of each pixel in the specific area to obtain some boundary lines based on the average grayscale value, wherein the grayscale values of pixels on the boundary lines are different from grayscale values of the other pixels which are not on the boundary lines; E) filtering horizontal and vertical lines of the boundary lines to obtain some remaining boundary lines; and F) fitting a defect specification line for the remaining boundary lines, wherein, if widths of the remaining boundary lines are not smaller than a width of the defect specification line, confirming that the remaining boundary lines are defect lines.

The invention relates to an AOI image scanning detection processing method. The method includes the following steps that: 1, an AOI device is provided with a CCD camera and performs line-by-line scanning on a detected circuit board through the CCD camera, so that a panoramic picture of the circuit board can be obtained; 2, the panoramic picture is subjected to vertical projection, so that the left shade, right shade and adhesion of a component can be eliminated; 3, the panoramic picture is subjected to horizontal projection, so that the front shade, back shade and adhesion of the component can be eliminated; 4, the panoramic picture is converted into a black-and-white picture through adopting a gray-scale critical value; and 5, binaryzation processing is performed on the black-and-white picture. According to the AOI image scanning detection processing method of the invention, shade and adhesion removal processing is performed on the picture before the picture is converted into the gray-scale picture, so that interferences and errors due to shades can be eliminated for the black-and-white picture after gray-scale conversion, and defect detection capability can be improved.

In a defect inspection apparatus (1), a processing image generation unit (61) calculates a feature quantity of each of pixels of two-dimensional image data, produced by an imaging device (5), by processing the two-dimensional image data with a predetermined algorithm. The processing image generation unit (61) extracts a pixel of a predetermined threshold value or more in the two-dimensional image data as a defect pixel. The processing image generation unit (61) generates processing image data such that the defect pixel is given a gradation value corresponding to the feature quantity while the remaining pixels other than the defect pixel are given zero gradation values. An analysis image generation unit (62) generates analysis image data consisting of one or a plurality of one-dimensional image data on the basis of the processing image data.

The invention relates to the field of image processing, in particular to a mechanical part surface defect detection method based on image processing, which comprises the following steps: acquiring a surface map and a grey-scale map of a to-be-detected mechanical part; carrying out canny edge detection on the grey-scale map to obtain a plurality of defect discrete edges and isolated points in the grey-scale map; determining an endpoint search domain and an intra-domain weight of each discrete edge according to the gradient amplitude and the gradient direction of each pixel point; obtaining the total similarity of isolated points in each discrete edge search domain; obtaining a connection coefficient corresponding to each isolated point according to the total similarity of the isolated points and the corresponding weight; carrying out endpoint updating on each discrete edge according to the connection coefficient; and performing stopping condition judgment on each discrete edge after the end points are updated each time, obtaining a final connection result, and then obtaining a defect area. The method is used for carrying out defect detection on the surface of the mechanical part, a more complete defect edge can be obtained through the method, and the defect detection effect is improved.

Methods and systems for enhanced defect detection based on images collected by at least two imaging detectors at different times are described. In some embodiments, the time between image measurements is at least 100 microseconds and no more than 10 milliseconds. In one aspect, one or more defects of interest are identified based on a composite image of a measured area generated based on a difference between collected images. In a further aspect, measurement conditions associated with the each imaged location are adjusted to be different for measurements performed by at least two imaging detectors at different times. In some embodiments, the measurement conditions are adjusted during the time between measurements by different imaging detectors. Exemplary changes of measurement conditions include environmental changes at the wafer under measurement and changes made to the optical configuration of the inspection system.

The invention discloses pipeline nondestructive testing equipment, and relates to the technical field of nondestructive testing, wherein the pipeline nondestructive testing equipment comprises a mobile carrier and a probe detection assembly; the mobile carrier moves in a pipeline along with fluid in the pipeline or moves through an advancing mechanism; the probe detection assembly comprises a detection part mounted on the mobile carrier, the detection part is provided with a detection end, and a detection probe is packaged in the detection end; and the pipeline nondestructive testing equipment further comprises a data processing unit, a first signal conditioning unit and a second signal conditioning unit, the detection probe comprises an excitation coil, a receiving coil and a passive resonance coil, and the passive resonance coil is located between the excitation coil and the receiving coil. Magnetization treatment equipment does not need to be additionally introduced during detection, so that the size of the system is greatly reduced, the requirement of the detection system on the internal cleanliness of the pipeline is reduced, the pipeline passing ability of the detection system is improved, and the cost overhead of the system is greatly reduced.